On Mon, Aug 17, 2026 at 10:17 AM Dongsheng Song via Gcc <[email protected]> wrote: > > On Sun, Aug 16, 2026 at 8:43 PM Zhongteng Gui via Gcc <[email protected]> wrote: > > > > Hi, I'm not sure whether this should be asked here, but I can't find a more > > appropriate list. > > > > I was reading https://gcc.gnu.org/gcc-17/criteria.html, and found that > > i686-mingw32 is listed as a secondary platform, while x86_64-w64-mingw32 is > > not > > listed. > > > > It seemes that i686-mingw32 has been here since gcc-5 (actually the primary > > platform list and secondary platform list only slightly changed from gcc-5 > > to > > gcc-17). However, i686-mingw32 itself is actually not a valid *triplet*. And > > regarding of i686-w64-mingw32, which it likely actually refers to, MSYS2 > > (the > > main distribution of GCC on Windows) has deprecated MINGW32 (the only > > environment with i686-w64-mingw32 triplet) since 2023-12-13 [1]. The default > > environment of MSYS2 (previously MINGW64 and now UCRT64) all uses > > x86_64-w64-mingw32. > > > > Considering that i686 devices are very rare compared to x86_64 nowadays, I > > think > > it would be better to switching from i686-mingw32 to x86_64-w64-mingw32 as a > > secondary platform. Maybe i686-pc-cygwin could also be replaced by > > x86_64-pc-cygwin. > > +1, all x86 platforms should transition to 64-bit. > > Same for freebsd, i586-unknown-freebsd -> x86_64-unknown-freebsd > (32-bit x86 unsupported since freebsd 15)
Maintainers of the respective ports should propose such changes, then they can easily be accepted. Richard.
